U.S. patent number 7,184,381 [Application Number 10/736,572] was granted by the patent office on 2007-02-27 for optical disk, optical disk recording/reproduction apparatus, and optical disk signal quality evaluation method.
This patent grant is currently assigned to Kabushiki Kaisha Toshiba, NEC Corporation. Invention is credited to Hiromi Honma, Toshiaki Iwanaga, Yutaka Kashihara, Yuuji Nagai, Masaki Nakano, Masatsugu Ogawa, Shuichi Ohkubo.
United States Patent |
7,184,381 |
Ohkubo , et al. |
February 27, 2007 |
Optical disk, optical disk recording/reproduction apparatus, and
optical disk signal quality evaluation method
Abstract
In order to evaluate the quality of a signal recorded on an
optical recording medium, a target signal is obtained based on a
predetermined data string and a predetermined partial response
characteristic, and for each clock cycle, an equalization error is
calculated that is a difference between the target signal and a
signal reproduced each clock cycle. Further, the product of the
equalization errors calculated at different times is obtained, and
based on the obtained product, the quality of a signal is
evaluated.
Inventors: |
Ohkubo; Shuichi (Tokyo,
JP), Honma; Hiromi (Tokyo, JP), Ogawa;
Masatsugu (Tokyo, JP), Nakano; Masaki (Tokyo,
JP), Iwanaga; Toshiaki (Tokyo, JP),
Kashihara; Yutaka (Yokohama, JP), Nagai; Yuuji
(Yokohama, JP) |
Assignee: |
NEC Corporation (Tokyo,
JP)
Kabushiki Kaisha Toshiba (Tokyo, JP)
|
Family
ID: |
32396328 |
Appl.
No.: |
10/736,572 |
Filed: |
December 17, 2003 |
Prior Publication Data
|
|
|
|
Document
Identifier |
Publication Date |
|
US 20040208101 A1 |
Oct 21, 2004 |
|
Foreign Application Priority Data
|
|
|
|
|
Dec 17, 2002 [JP] |
|
|
2002/365772 |
Nov 17, 2003 [JP] |
|
|
2003/386726 |
|
Current U.S.
Class: |
369/59.22;
369/124.01; 369/53.1; 369/59.19; G9B/20.01; G9B/27.052 |
Current CPC
Class: |
G11B
20/10009 (20130101); G11B 20/10055 (20130101); G11B
20/1012 (20130101); G11B 20/10296 (20130101); G11B
2220/20 (20130101); G11B 2220/216 (20130101); G11B
2220/218 (20130101) |
Current International
Class: |
G11B
5/09 (20060101) |
References Cited
[Referenced By]
U.S. Patent Documents
Primary Examiner: Edun; Muhammad
Attorney, Agent or Firm: Mion, PLLC; Sughrue
Claims
What is claimed is:
1. A signal quality evaluation method for a reproduced equalized
signal obtained by reproducing and equalizing a signal recorded on
an optical disk medium by using embossing, or by using an optical
data recording apparatus, comprising the steps of: calculating, for
each clock cycle, an equalization error between a target signal,
which is obtained based on a predetermined data string and a
predetermined partial response characteristic, and a signal
reproduced each clock cycle; and evaluating a signal quality based
on the auto-correlation of the equalization error.
2. A signal quality evaluation method according to claim 1, wherein
h.sub.0=1, h.sub.1=2, h.sub.2=2, h.sub.3=2 and h.sub.4=1 are used
as the partial response characteristic and
R.sub.1=.SIGMA.v.sub.kv.sub.k+1/N is determined, wherein v.sub.k
denotes an equalization error and N denotes the number of samples;
and wherein the quality of the signal recorded on the optical disk
is evaluated by examining a first signal quality evaluation value
S.sub.1, a second signal quality evaluation value S.sub.2 and a
third signal quality evaluation value S.sub.3 that are represented
by the following equations (20), (21) and (22)
.times..times..times..times..times..times..times..times..times..times..ti-
mes..times. ##EQU00012##
3. An optical disk, on which data are recorded under a condition
wherein the first signal quality evaluation value S.sub.1, the
second signal quality evaluation value S.sub.2 and the third signal
quality evaluation value S.sub.3 according to claim 2 are equal to
or greater than 12.
4. An optical disk, on which data re recorded under a condition
wherein the first signal quality evaluation value S.sub.1, the
second signal quality evaluation value S.sub.2 and the third signal
quality evaluation value S.sub.3 according to claim 2 are equal to
or greater than 15.
5. An optical disk recording/reproduction apparatus or reproduction
apparatus for performing data recording or data reproduction under
a condition wherein the first signal quality evaluation value
S.sub.1, the second signal quality evaluation value S.sub.2 and the
third signal quality evaluation value S.sub.3 according to claim 2
are equal to or greater than 12.
6. An optical disk recording/reproduction apparatus or reproduction
apparatus for performing data recording or data reproduction under
a condition wherein the first signal quality evaluation value
S.sub.1, the second signal quality evaluation value S.sub.2 and the
third signal quality evaluation value S.sub.3 according to claim 2
are equal to or greater than 15.
7. A signal quality evaluation method for a reproduced equalized
signal obtained by reproducing and equalizing a signal recorded on
an optical disk medium by using embossing, or by using an optical
data recording apparatus, comprising the steps of: projecting an
equalization error onto a noise vector that is defined by using a
partial response characteristic and a difference between two sets
of time series data; and evaluating a signal quality based on a
ratio of the variance of the projected equalization errors to an
Euclid distance that is defined by using the partial response
characteristic and the difference between the two sets of time
series data.
8. An optical disk signal quality evaluation method for a
reproduced equalized signal obtained by reproducing and equalizing
a signal recorded on an optical disk medium by using embossing, or
by using an optical data recording apparatus, comprising the steps
of: designating an equalization error
v.sub.k=(y.sub.k-.SIGMA.a.sub.k-ih.sub.i), wherein y.sub.k denotes
the value of a signal reproduced and equalized for each clock
cycle, a.sub.k denotes a predetermined data string and h.sub.k
denotes a predetermined partial response characteristic, and
designating a time delay operator D that uses a clock time as a
unit; designating as a polynomial A(D)=.SIGMA..alpha..sub.jD.sub.j,
which is defined by using .alpha..sub.j, a coefficient of either 1,
0 or -1, and which satisfies .alpha..sub.j.alpha..sub.j+1.noteq.-1,
and designating H(D)=.SIGMA.h.sub.kD.sub.k as a PR polynomial that
defines a partial response; based on a polynomial defined as
N(D)=A(D)H(D)=.SIGMA..epsilon..sub.iD.sub.i, calculating a signal
quality evaluation value S that is defined by the following
equation (18) .function..times..times..times..times. ##EQU00013##
and evaluating the quality of a signal recorded on an optical
disk.
9. A signal quality evaluation method according to claim 8, wherein
the quality of the signal recorded on the optical disk is evaluated
based on the signal quality evaluation value S obtained for at
least two sets in sets of tap coefficients .epsilon..sub.i, one of
which is a set of tap coefficients .epsilon. that provide the
smallest Euclid distance d=.SIGMA..epsilon..sub.i.sup.2 and another
one of which is a set of tap coefficients .epsilon. that provide
the second smallest Euclid distance d.
10. A signal quality evaluation method according to claim 8,
wherein h.sub.0=1, h.sub.1=2, h.sub.2=2, h.sub.3=2 and h.sub.4=1
are used as the partial response characteristic, and the quality of
the signal recorded on the optical disk is evaluated based on the
signal quality evaluation value S that is obtained for each set of
tap coefficients .epsilon. that provide an Euclid distance d of 12
or 14.
11. A signal quality evaluation method according to claim 8,
wherein h.sub.0=1, h.sub.1=2, h.sub.2=2, h.sub.3=2 and h.sub.4=1
are used as the partial response characteristic, and the quality of
the signal recorded on the optical disk is evaluated based on the
signal quality evaluation value S that is obtained for each of at
least three sets of tap coefficients .epsilon. that are represented
by the following equation (19) .epsilon.:.epsilon..sub.0=1,
.epsilon..sub.1=2, .epsilon..sub.2=2, .epsilon..sub.3=2,
.epsilon..sub.4=1 .epsilon.:.epsilon..sub.0=1, .epsilon..sub.1=2,
.epsilon..sub.2=1, .epsilon..sub.3=0, .epsilon..sub.4=-1,
.epsilon..sub.5=2, .epsilon..sub.6=-1 .epsilon.:.epsilon..sub.0=1,
.epsilon..sub.1=2, .epsilon..sub.2=1, .epsilon..sub.3=0,
.epsilon..sub.4=0, .epsilon..sub.5=0, .epsilon..sub.6=1,
.epsilon..sub.2=2, .epsilon..sub.8=1 (19).
12. An optical disk, on which data are recorded under a condition
wherein the signal quality evaluation value S according to claim 8
is equal to or greater than 12.
13. An optical disk, on which data are recorded under a condition
wherein the signal quality evaluation value S according to claim 8
is equal to or greater than 15.
14. An optical disk recording/reproduction apparatus or
reproduction apparatus for performing data recording or data
reproduction under a condition wherein the signal quality
evaluation value S according to claim 8 is equal to or greater than
12.
15. An optical disk recording/reproduction apparatus or
reproduction apparatus for performing data recording or data
reproduction under a condition wherein the signal quality
evaluation value S according to claim 8 is equal to or greater than
15.
16. A signal quality evaluation method according to one of claims 1
to 8, wherein the predetermined data string is binary data for the
reproduced equalized signal obtained by a Viterbi decoder.
17. A signal quality evaluation method according to claim 16,
wherein the number N of samples is equal to or greater than
10.sup.4.
18. A signal quality evaluation method according to one of claims 8
to 2, whereby wherein the number N of samples is equal to or
greater than 10.sup.4.
19. A signal quality evaluation apparatus, for a reproduced
equalized signal y.sub.k that is obtained by reproducing and
equalizing a signal that has been recorded in advance on an optical
disk medium by using embossing or by using an optical data
recording apparatus, comprising: a target signal generator for
generating a target signal .SIGMA.a.sub.k-ih.sub.i based on a
predetermined data string a.sub.k and a predetermined partial
response characteristic h.sub.k; a computation unit for using the
reproduced equalized signal y.sub.k and the target signal
.SIGMA.a.sub.k-ih.sub.i to calculate an equalization error
v.sub.k=(y.sub.k-.SIGMA.a.sub.k-ih.sub.i); and means for using
auto-correlation for the equalization error to calculate a quality
evaluation value for the reproduced equalized signal y.sub.k.
20. A signal quality evaluation apparatus, for a reproduced
equalized signal y.sub.k that is obtained by reproducing and
equalizing a signal that has been recorded in advance on an optical
disk medium by using embossing or by using an optical data
recording apparatus, comprising: a target signal generator for
generating a target signal .SIGMA.a.sub.k-ih.sub.i based on a
predetermined data string a.sub.k and a predetermined partial
response characteristic h.sub.k; a computation unit for using the
reproduced equalized signal y.sub.k and the target signal
.SIGMA.a.sub.k-ih.sub.i to calculate an equalization error
v.sub.k=(y.sub.k-.SIGMA.a.sub.k-ih.sub.i); a delay element group,
including a plurality of delay elements, for receiving the
equalization error, and for outputting equalization errors v.sub.k,
v.sub.k-1, . . . and v.sub.k-1+1 a plurality of times; means for
receiving the equalization errors v.sub.k, v.sub.k-1, . . . and
v.sub.k-n+1 and for calculating R.sub.n(n=0, 1, 2, . . . to L-1),
based on the following equation (23), and outputting R.sub.n; noise
variance calculation means for performing weighting for R.sub.n
with a coefficient .beta..sub.0, .beta..sub.1, . . . or
.beta..sub.L-1 to obtain a noise variance
.SIGMA..beta..sub.iR.sub.i and quality evaluation value calculation
means for using the noise variance .SIGMA..beta..sub.iR.sub.i to
calculate a quality evaluation value for the reproduced equalized
signal, wherein Rn=E {V.sub.k, v.sub.k-n}, and E {x.sub.i, y.sub.j}
is the average value of the product x.sub.iy.sub.i (23).
21. A signal quality evaluation apparatus, for a reproduced
equalized signal y.sub.k that is obtained by reproducing and
equalizing a signal that has been recorded in advance on an optical
disk medium by using embossing or by using an optical data
recording apparatus, comprising: a target signal generator for
generating a target signal .SIGMA.a.sub.k-ih.sub.i based on a
predetermined data string a.sub.k and a predetermined partial
response characteristic (h.sub.0=1, h.sub.1=2, h.sub.2=2,
h.sub.3=2, h.sub.4=1); a computation unit for using the reproduced
equalized signal y.sub.k and the target signal
.SIGMA.a.sub.k-ih.sub.i to calculate an equalization error
v.sub.k=(y.sub.k-.SIGMA.a.sub.k-ih.sub.i); a delay element group,
including a plurality of delay elements, for receiving the
equalization error, and for outputting equalization errors v.sub.k,
v.sub.k-1, . . . and v.sub.k-n+1 a plurality of times; means for
receiving the equalization errors v.sub.k, v.sub.k-1, . . . and
v.sub.k-n+1, and for calculating R.sub.n(n=0, 1, 2, . . . to L-1),
based on the following equation (24), and outputting R.sub.n; noise
variance calculation means for performing weighting for R.sub.n
with a coefficient .beta..sub.0, .beta..sub.1, . . . or
.beta..sub.L-1 to obtain a noise variance
.SIGMA..beta..sub.iR.sub.i, and quality evaluation value
calculation means for using the noise variance
.SIGMA..beta..sub.iR.sub.i to calculate a quality evaluation value
for the reproduced equalized signal, wherein Rn=E {v.sub.k,
v.sub.k-n}, and .SIGMA.{x.sub.i, y.sub.i} is the average value of
the product x.sub.iy.sub.i (24).
22. A signal quality evaluation apparatus according to claim 12,
wherein the noise variance calculation means calculates a first
noise variance .sigma..sub.1.sup.2 using the coefficients
.beta..sub.0=1, .beta..sub.1=12/7, .beta..sub.2=8/7,
.beta..sub.3=4/7 and .beta..sub.4=1/7, calculates a second noise
variance .sigma..sub.2.sup.2 using the coefficients .beta..sub.0=1,
.beta..sub.1=8/6, .beta..sub.2=1/6, .beta..sub.3=-4/6,
.beta..sub.4=-1, .beta..sub.5=-4/6 and .beta..sub.6=-1/6, and
calculates a third noise variance .sigma..sub.3.sup.2 using the
coefficients .beta..sub.0=1, .beta..sub.1=8/6, .beta..sub.2=2/6,
.beta..sub.3=0, .beta..sub.4=1/6, .beta..sub.5=4/6, .beta..sub.6=1,
.beta..sub.7=4/6 and .beta..sub.8=1/6; and wherein the quality
evaluation value calculation means calculates the quality
evaluation value by using the smallest of the values
(14/.sigma..sub.1.sup.2), (12/.sigma..sub.2.sup.2) and
(12/.sigma..sub.3.sup.2).
23. An optical disk apparatus on which a quality evaluation
apparatus according to one of claims 19 to 22 is mounted.
24. A signal quality evaluation apparatus according to one of
claims 10 to 22, wherein the predetermined data string is obtained
by performing Viterbi decoding for the reproduced equalized
signal.
25. An optical disk recording/reproduction or reproduction
apparatus comprising: means for generating a target signal based on
the value of a signal reproduced for each clock cycle, a
predetermined data string and a predetermined partial response
characteristic; and means for calculating an equalization error
that is the difference between the signal reproduced for each clock
cycle and the target signal.
26. An optical disk recording/reproduction or reproduction
apparatus according to claim 25, further comprising: means for
performing addition or multiplication, or a sum of product
operation, for equalization errors occurring at different
times.
27. An optical disk recording/reproduction or reproduction
apparatus according to claim 25 or 26, wherein at least 10.sup.4
equalization errors are obtained through calculation.
Description
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to an optical disk, an optical disk
recording/reproduction apparatus, and a method for evaluating
optical data of the optical disk.
2. Related Background Art
Optical disks are classified into two types: read-only optical
disks on which data are recorded in advance and optical disks on
which data can be recorded. To record data on the read-only optical
disk, generally, an exposure process called a mastering process is
performed to form embossments (physically raised and recessed
conditions) on the optical disk. Then, to record data on the
recordable optical disk, the optical disk is irradiated with a
focused laser beam to change a specific physical characteristic of
a recording film.
Conventionally, when evaluating the quality of data or signals
recorded on an optical disk, a measurement is generally made of the
jitter characteristic of a reproduction signal that is obtained by
irradiating with a laser beam and that is reflected by the optical
disk. As is shown in FIG. 1, the jitter characteristic is usually
represented as the timetransient fluctuation of an edge position
that is obtained by slicing a reproduction signal at a specific
reference potential.
However, as is apparent from FIG. 1, when the recording density is
increased, and when a mark length is small, the signal amplitude is
reduced so it does not extend across the slice level used for
detecting the edge position. Therefore, performing the jitter
measurement is difficult. Further, as is shown in FIG. 1, when the
recording density is further increased to improve the recording
capacity, the effect produced by intersymbol interference is
increased and causes the signal amplitude to be reduced so that the
signal amplitude does not extend across the slice level used for
the detection of the edge position. As a result, performing the
jitter measurement is difficult.
Conventionally, to reduce the intersymbol interference, a method
has been employed that uses an equalizer to filter a reproduced
waveform. However, while restricting the intersymbol interference,
the equalizer generally increases the noise component. Thus, when
the recording density is very high, it is difficult for the
original recorded data to be obtained by decoding the reproduced
signal.
As a method for accurately decoding data when the recording density
is very high, there is a well known signal detection method called
the PRML (Partial-Response Maximum-Likelihood) method. According to
this method, a reproduced waveform is equalized (PR equalized) to
provide a waveform that includes intersymbol interference for
suppressing the noise component, and the data are identified by
employing a method called Viterbi decoding (ML). In this instance,
the PR equalization is defined by the amplitude for each data cycle
(clock). For example, for PR(abc), the amplitude at time 0 is a,
the amplitude at time T is b, the amplitude at time 2T is c and the
amplitude at all other times is 0. The total number of components
having an amplitude other than 0 is called a constraint length.
According to the PRML, instead of detecting the edge position to
decode data, the value obtained by sampling a reproduced waveform
each clock cycle is employed to obtain data through the Viterbi
decoding. Therefore, it is difficult to estimate the detection
function of the PRML based only on the time-transient fluctuation
data for the edge position.
SUMMARY OF THE INVENTION
It is, therefore, one object of the present invention to provide a
signal quality evaluation index to be used instead of jitter when
the density is too high to measure the jitter, and a reference
value therefor.
According to one aspect of the present invention, provided is a
quality evaluation method for a reproduced equalized signal
obtained by reproducing and equalizing a signal recorded on an
optical disk by using embossing, or by using an optical data
recording apparatus. This method of the present invention comprises
the steps of:
calculating, for each clock cycle, an equalization error between a
target signal, which is obtained based on a predetermined data
string and a predetermined partial response characteristic, and a
signal reproduced each clock cycle; and
evaluating a signal quality based on the auto-correlation of the
equalization error.
Further, according to another aspect of the invention, a signal
quality evaluation method comprises the steps of:
projecting an equalization error onto a noise vector that is
defined by using a partial response characteristic and a difference
between two sets of time series data; and
evaluating a signal quality based on a ratio of the variance of the
equalization errors to an Euclid distance that is defined by using
the partial response characteristic and the difference between the
two sets of time series data.
According to a third aspect of the present invention, a signal
quality evaluation method comprises the steps of:
designating an equalization error
v.sub.k=(y.sub.k-.SIGMA.a.sub.k-ih.sub.i), wherein y.sub.k denotes
the value of a signal reproduced and equalized for each clock
cycle, a.sub.k denotes a predetermined data string and h.sub.k
denotes a predetermined partial response characteristic, and
designating a time delay operator D that uses a clock time as a
unit;
designating as a polynomial A(D)=.alpha..sub.1D.sub.1, which is
defined by using .alpha..sub.1, a coefficient of either 1, 0 or -1,
and which satisfies .alpha..sub.j.alpha..sub.j+1.noteq.-1, and
designating H(D)=h.sub.kD.sub.k as a PR polynomial that defines a
partial response;
based on a polynomial defined as
N(D)=A(D)H(D)=.SIGMA..epsilon..sub.iD.sub.i, calculating a signal
quality evaluation value S that is defined by the following
equation (1)
.times..times..times..times..times..times..times. ##EQU00001##
and
evaluating the quality of a signal recorded on an optical disk.
In addition, the quality of the signal recorded on the optical disk
is evaluated based on the signal quality evaluation value S
obtained for at least two sets in sets of tap coefficients
.epsilon..sub.i, one of which is a set of tap coefficients
.epsilon. that provide the smallest Euclid distance
d=.SIGMA..epsilon..sub.i.sup.2 and another one of which is a set of
tap coefficients .epsilon. that provide the second smallest Euclid
distance d.
Furthermore, h.sub.0=1, h.sub.1=2, h.sub.2=2, h.sub.3=2 and
h.sub.4=1 are used as the partial response characteristic, and the
quality of the signal recorded on the optical disk is evaluated
based on the signal quality evaluation value S that is obtained for
each set of tap coefficients a that provide an Euclid distance d of
12 or 14.
Further, h.sub.0=1, h.sub.1=2, h.sub.2=2, h.sub.3=2 and h.sub.4=1
are used as the partial response characteristic, and the quality of
the signal recorded on the optical disk is evaluated based on the
signal quality evaluation value S that is obtained for each of at
least three sets of tap coefficients .epsilon. that are represented
by the following equation (2). .epsilon.:.epsilon..sub.0=1,
.epsilon..sub.1=2, .epsilon..sub.2=2, .epsilon..sub.3=2,
.epsilon..sub.4=1 .epsilon.:.epsilon..sub.0=1, .epsilon..sub.1=2,
.epsilon..sub.2=1, .epsilon..sub.3=0, .epsilon..sub.4=-1,
.epsilon..sub.5=-2, .epsilon..sub.6=-1 .epsilon.:.epsilon..sub.0=1,
.epsilon..sub.1=2, .epsilon..sub.3=0, .epsilon..sub.4=0,
.epsilon..sub.5=0, .epsilon..sub.6=1, .epsilon..sub.7=2,
.epsilon..sub.8=1 (2)
Moreover, h.sub.0=1, h.sub.1=2, h.sub.2=2, h.sub.3=2 and h.sub.4=1
are used as the partial response characteristic and
R.sub.i=.SIGMA.v.sub.kv.sub.k+1/N is determined, wherein v.sub.k
denotes an equalization error and N denotes the number of samples.
And the quality of the signal recorded on the optical disk is
evaluated by examining a first signal quality evaluation value
S.sub.1, a second signal quality evaluation value S.sub.2 and a
third signal quality evaluation value S.sub.3 that are represented
by the following equations (3), (4) and (5).
.times..times..times..times..times..times..times..times..times..times..ti-
mes..times. ##EQU00002##
In addition, according to the present invention, binary data
obtained by a Viterbi decoder are used for a data string used to
generate a target signal for calculating the equalization error.
Further, the signal quality is evaluated by using 10.sup.4, or
more, equalization errors.
Furthermore, according to the present invention, an optical disk is
provided on which data are recorded under a condition wherein the
signal quality evaluation value S, or the first signal quality
evaluation value S.sub.1, the second signal quality evaluation
value S.sub.2 and the third signal quality evaluation value
S.sub.3, are equal to or greater than 12, and preferably are equal
to or greater than 15.
According to the present invention, an optical disk
recording/reproduction or reproduction apparatus performs data
recording or data reproduction under a condition wherein the signal
quality evaluation value S, or the first signal quality evaluation
value S.sub.1, the second signal quality evaluation value S.sub.2
and the third signal quality evaluation value S.sub.3, are equal to
or greater than 12, and preferably are equal to or greater than
15.
According to the present invention, an optical disk
recording/reproduction or reproduction apparatus comprises:
a function for generating a target signal based on the value of a
signal reproduced for each clock cycle, a predetermined data string
and a predetermined partial response characteristic; and
a function for calculating an equalization error that is the
difference between the signal reproduced for each clock cycle and
the target signal.
Furthermore, according to the present invention, an optical disk
recording/reproduction or reproduction apparatus comprises:
a function for performing addition or multiplication, or a sum of
product operation, for equalization errors occurring at different
times. In addition, at least 10.sup.4 equalization errors are
obtained through calculation,
For the PRML, an algorithm called Viterbi decoding is employed to
discriminate data. According to the Viterbi decoding, the square of
a difference between the value of a reproduction signal and a
predetermined level defined by a partial response is calculated,
the square sum is obtained along each path, and a path providing
the smallest square sum is selected to decode the data.
When the Euclid distance between the paths is small, a detection
error tends to occur in the Viterbi decoding. The Euclid distance d
between different paths is defined as
d.sup.2=4.SIGMA..epsilon..sub.i.sup.2 when
B(D)=.SIGMA.b.sub.kD.sub.k denotes a polynomial defined based on a
data string b.sub.k along one of the paths,
C(D)=.SIGMA.c.sub.kD.sub.k denotes a polynomial defined based on a
data string C.sub.k along the other path (b.sub.k and c.sub.k are
binary data of 1 or -1), and
N(D)=(B(D)-C(D))H(D)=2.SIGMA..epsilon..sub.iD.sub.i is established,
wherein H(D)=.SIGMA.h.sub.kD.sub.k is a polynomial defining a
partial response. In this case, D denotes a time delay operator
using a clock time as a unit, and h.sub.k denotes a predetermined
partial response characteristic. The partial response
characteristic is represented as PR(h.sub.0, h.sub.1, h.sub.2,
h.sub.3 . . . ) generally by using the element of h.sub.k that is
not 0.
Assume that the partial response characteristic is defined as
h.sub.0=1, h.sub.1=2 and h.sub.2=1, while h.sub.3 and the
following=0; that the data string b.sub.k is defined as b.sub.0=1,
b.sub.1=1 and b.sub.2=-1, while b.sub.3 and the following=-1; and
that the data string c.sub.k is defined as c.sub.0=-1, c.sub.1=1
and c.sub.2=1, while c.sub.3 and the following=-1. In this case,
since
N(D)=2(1-D.sup.2)(1+2D+D.sup.2)=2.times.(1+2D-2D.sup.3-D.sup.4),
the Euclid distance d between the path along the data string
b.sub.k and the path along the data string c.sub.k is obtained as
d.sup.2=4.times.(1.times.1+2.times.2+2.times.2+1.times.1).
To express the binary data, a combination of 1 and 0, or a
combination of 1 and -1, is used, and in this invention, the
combination of 1 and -1 is used. Since the amplitude of a waveform
obtained when the combination of 1 and 0 is used is 1/2 the
amplitude of a waveform when the combination of 1 and -1 is used,
the numerators for the evaluation values S.sub.1 to S.sub.3 defined
by equations 10 to 12 need only be set to 1/4. That is, only 3.5 or
3 need be used.
When the PR polynomial is defined, the Euclid distance d between
the paths can be calculated for each set of tap coefficients
.epsilon..sub.1. For an optical disk, generally, a record symbol
d>1 is used to limit the run length, and when, for example, the
recording symbol is d=1, a mark having a length equal to or greater
than 2T is recorded on the disk. In order to take this limitation
into account for the calculation of the Euclid distance,
restriction .epsilon..sub.i.epsilon..sub.j+1.noteq.-1 need only be
provided for the set of tap coefficients .epsilon..sub.i. That is,
(x, 1, -1, y), for example, can be used as the data string b.sub.k
that satisfies .epsilon..sub.i.epsilon..sub.j+1, and (x, 01, 1, y)
can be used as the data string c.sub.k. However, since the pattern
(1, -1, 1) or (-1, 1, -1) is prohibited by the limitation d=1, with
x=-1 or y=-1, the data string b.sub.k becomes a pattern that do s
not conform to the run length limitation (a pattern that can not
exist), and with x=1 or y=-1, the data string c.sub.k is a pattern
that is not compatible with the run length limitation. Therefore, a
combination of the data strings b.sub.k and c.sub.k does not exist
that satisfies .epsilon..sub.i.epsilon..sub.i+1=-1, while the run
length limitation is satisfied. Further, when the length of a mark
recorded on the disk is equal to or greater than 3T, only the
restriction imposed by .epsilon..sub.i.epsilon..sub.i+1.noteq.-1
and .epsilon..sub.i.epsilon..sub.i+2.noteq.-1 need be provided.
While, for example, the data string b.sub.k is used as a reference,
the probability whereat a detection error will occur for the Euclid
distance d between the two paths is equivalent to the probability
whereat .SIGMA.(y.sub.k-.SIGMA.b.sub.k-ih.sub.i).sup.2 will be
greater than .SIGMA.(y.sub.k-.SIGMA.c.sub.k-ih.sub.i).sup.2 due to
noise. When the data string b.sub.k is used as a reference,
y.sub.k-.SIGMA.b.sub.k-ih.sub.i is an equalization error. Further,
the difference between
.SIGMA.(y.sub.k-.SIGMA.b.sub.k-ih.sub.i).sup.2 and
.SIGMA.(y.sub.k-.SIGMA.c.sub.k-ih.sub.i).sup.2 may be observed
using the following method. An error vector is defined by
regarding, as the elements of a vector, the coefficients of the
polynomial defined by using the difference between B(D)H(D) and
C(D)H(D), and the equalization error is projected onto the error
vector. In this case, the probability of the occurrence of a
detection error is defined as the probability whereat the magnitude
of the noise (the variance of the noise) projected onto the error
vector is greater than half the Euclid distance between the paths.
Therefore, when the ratio of the Euclid distance between the paths
to the variance of the noise projected onto the error vector need
only be calculated to estimate the quality of a signal. When data
are obtained in advance when the recording condition is adjusted,
this data string can be used as a reference data string, and when
such data have not yet been obtained, probable binary data that are
obtained by a Viterbi decoder can be used.
When the data string b.sub.k is defined as b.sub.0=-1 and b.sub.1=1
and b.sub.2 and the following=1, and when the data string c.sub.k
is defined as c.sub.0=1 and c.sub.1 and the following=1,
.alpha..sub.0=0 and .alpha..sub.1 and the following=0 are obtained
while A(D)=C(D)-B(D)=2.SIGMA..alpha..sub.jD.sub.j. When, for
example, h.sub.0=1, h.sub.1=2, h.sub.2=2 and h.sub.3=1 are used as
H(D) (corresponding to PR(1,2,2,1)), the coefficients
.epsilon..sub.i of polynomial
N(D)=A(D)H(D)=2.SIGMA..epsilon..sub.iD.sub.i, which defines the
error vector, are (1,2,2,1) in the order .epsilon..sub.0,
.epsilon..sub.1, .epsilon..sub.2 and .epsilon..sub.3. Therefore,
the probability whereat the data string b.sub.k will be erroneously
regarded as the data string c.sub.k for PR(1221) equals the
probability whereat the magnitude of the equalization error
projected onto 2.times.(1,2,2,1) is greater than half the Euclid
distance (in this case, 2.times.(1+2.times.2+2.times.2+1).sup.1/2)
between the two paths. Since the projection of the equalization
error onto the error vector is represented by the following
equation (6),
.times..times..times..times..times..times..times. ##EQU00003## the
variance CN of the noise projected onto the error vector is
represented by the following equation (7).
.times..times..times..times..times..times..times..times.
##EQU00004## Further, since half of the Euclid distance between the
two paths, which corresponds to the signal amplitude, is
represented by the following equation (8),
.times..times. ##EQU00005## and since the square E of the amplitude
that corresponds to electric power is represented by the following
equation (9),
.times..times. ##EQU00006## E/CN can be obtained as an index that
is correlated with the error probability.
Since coefficient 2 related to the entire A(D) and N(D) does not
affect the calculation results, the same results are obtained
through the calculation using A(D)=.SIGMA..alpha..sub.iD.sub.i and
N(D)=.SIGMA..epsilon..sub.iD.sub.i, with the coefficient 2 being
omitted.
As is described above, v.sub.k=(y.sub.k-.SIGMA.a.sub.k-ih.sub.i) is
defined as an equalization error wherein y.sub.k denotes the value
of a signal reproduced and equalized for each clock cycle, a.sub.k
denotes a predetermined data string for generation of a target
signal and h.sub.k denotes a predetermined partial response, while
D is defined as a time delay operator using a clock time as a unit.
Further, A(D)=.SIGMA..alpha..sub.jD.sub.j is a polynomial, defined
by using a coefficient .alpha..sub.j having a value either of 1, 0
or -1, that satisfies .alpha..sub.j.alpha..sub.j+1.noteq.-1, and
H(D)=.SIGMA.h.sub.kD.sub.k is a PR polynomial that defines a
partial response. Then, when the polynomial defined as
N(D)=A(D)H(D)-.SIGMA..sub.jD.sub.j is employed to calculate the
signal quality evaluation value defined by the following equation
(10),
.times..times..times..times..times..times..times. ##EQU00007## the
probability that a detection error will occur can be obtained,
i.e., the quality of a reproduced signal can be evaluated.
In the above explanation, the data strings b.sub.k and c.sub.k have
been used as an example combination of data strings that tend to be
erroneously regarded. However, when the variance of the noise
projected onto the error vector is to be calculated, specific data
strings need not always be selected to obtain an equalization
error. That is, for the calculation of the equalization error
variance, a time corresponding to the data string b.sub.k need not
be extracted from the data string a.sub.k used for the generation
of a target signal. Instead, the equalization error obtained for
each clock time can be used to calculate the variance. This is
because, so long as the equalization errors are stochastically
distributed in accordance with the Gaussian distribution, the same
results are obtained either by extracting a specific portion and
calculating the variance, or by using the entire portion and
calculating the variance. The variance of the noise may be
calculated by using only a specific data string b.sub.k; however,
when the variance of the equalization errors is calculated without
selecting a pattern, the configuration of the circuit is
simplified.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is a schematic graph for explaining change of a reproduced
waveform when a recording density is changed;
FIG. 2 is a diagram showing a configuration example for a
functional block used to measure signal quality;
FIG. 3 is a diagram showing a configuration example for a signal
quality evaluation unit for calculating the variance of
equalization errors projected onto an error vector;
FIG. 4 is a graph showing the relationship between a signal quality
evaluation value S and a bit error rate bER;
FIG. 5 is a graph showing the relationship between the signal
quality evaluation value S and the bit error rate bER;
FIG. 6 is a diagram showing another configuration example for the
signal quality evaluation unit for calculating the signal quality
evaluation value S;
FIG. 7 is a graph showing the relationship between the number of
samples and the signal quality evaluation value S;
FIG. 8 is a diagram showing a configuration example for a
functional block for calculating the signal quality evaluation
value S when recorded data are already known;
FIG. 9 is a diagram showing an optical data recording/reproduction
apparatus according to the present invention that has a function
for adjusting a recording condition or a reproduction condition;
and
FIG. 10 is a diagram showing a signal quality evaluation unit for
calculating the signal equality evaluation value S by using only a
target signal that has been predetermined.
DESCRIPTION OF THE PREFERRED EMBODIMENTS
FIG. 2 is a diagram showing an example of functional block for
calculating a signal quality evaluation value S. An A/D converter
10 performs sampling for a reproduced signal having a constant
frequency, and an equalizer 11 that includes a PLL (phase lock
loop) circuit obtains equalized reproduced waveform data each clock
cycle. In the equalization process, while the noise component is
suppressed to the extent possible, the reproduced waveform is
equalized so that it is as similar as possible to a target waveform
based on a PR waveform. A signal quality evaluation unit 12
calculates the signal quality evaluation value S by using the
received, equalized reproduced waveform, and evaluates the quality
of the reproduced waveform. When a read-only optical disk or a
recordable optical disk on which data are recorded by another
recording apparatus is employed, original data a.sub.k recorded on
the disk are not always known in advance. In this case, binary data
obtained by a discriminator (typically, a Viterbi decoder) included
in the signal quality evaluation unit 12 may be employed as the
data a.sub.k.
FIG. 3 is a detailed diagram showing the signal quality evaluation
unit 12 in FIG. 2. When a reference PR waveform is denoted by
h.sub.1, and binary data obtained by a discriminator 20 (typically,
a Viterbi decoder) is denoted by a.sub.k, a target signal generator
21 generates a target signal R.sub.k based on the following
equation (11).
.times..times. ##EQU00008##
Then, a comparator 22 calculates an equalization error v.sub.k that
is the difference between a signal y.sub.k, reproduced (equalized)
for each clock cycle, and the target signal R.sub.k. According to
this configuration, in order to project the equalization error
v.sub.k onto an error vector, the equalization error v.sub.k is
delayed for each clock cycle by a plurality of taps 23, and the
results are added by an adder 23-2 through a plurality of
coefficient multipliers 23-1 each of which is for multiplying by
tap coefficient .epsilon..sub.i. In this configuration, the
plurality of taps, namely, unit delay elements 23, the plurality of
coefficient multipliers 23-1 and one adder 23-2 constitute a
transversal filter TRF.
The tap coefficient .epsilon..sub.i is a coefficient of the N(D)
described above. When each set of tap coefficients .epsilon..sub.i
is identified by using "j", and a corresponding signal quality
evaluation value S is defined as a j-th signal quality evaluation
value S.sub.j, a total number m of the taps 23(D) in FIG. 3 is
changed, depending on j. Assume that a set of tap coefficients
.epsilon..sub.j relative to j=1 is .epsilon..sub.0=1,
.epsilon..sub.1=1, .epsilon..sub.2=0, .epsilon..sub.3=0,
.epsilon..sub.4=1 and .epsilon..sub.5 and the following =0, and
that a set of tap coefficients .epsilon..sub.j relative to j=2 is
.epsilon..sub.0=1, .epsilon..sub.1=2, .epsilon..sub.2=1 and
.epsilon..sub.3 and the following=0. In this case, m=4 is obtained
to calculate the first signal quality evaluation value S.sub.1, and
m=2 is obtained to calculate the second signal quality evaluation
value S.sub.2. When changing of the number of taps in accordance
with the combination j is complex, only a satisfactorily large
number of taps (e.g., 15 to 20) need be prepared, and only the tap
coefficients that are not required for the calculation need be set
to 0.
Following this, a multiplier 24 calculates the square of
equalization error .SIGMA..epsilon..sub.i.times.v.sub.k+1, which is
obtained through the plurality of tap coefficient multipliers 23-1
and the adder 23-2, and an adder 25 multiplies the resultant square
by the total number N of equalization error samples. As a result, a
value is obtained that is proportional to the variation of the
noise projected onto the error vector. When this value is divided
by N and .SIGMA..epsilon..sub.i.sup.2, a variation value is
obtained, and since N and .SIGMA..epsilon..sub.i.sup.2 are
constants, these are employed as the coefficients by a multiplier
27. A divider 26 calculates the reciprocal of the obtained value,
namely, the output of the adder 25, and the multiplier 27
calculates a product
(N.SIGMA..epsilon..sub.i.sup.2).times..SIGMA..epsilon..sub.i.sup.2.
As a result, the signal quality evaluation value S is obtained.
Actually, (N+m) reproduced waveform samples are required; however,
since m is at most 20 while N is equal to or greater than 10.sup.4,
in this invention, N is consistently employed as the number of
samples. The required total number of samples will be explained
later.
[First Embodiment]
The relationship between a bit error rate (bER) and the signal
quality evaluation value S is examined by a read-only optical disk
(ROM disk), provided in advance by using embossing to record a pit
data string on a polycarbonate substrate that is 0.6 mm thick.
In this embodiment, the pit string was formed on the substrate as
random data that was modulated based on the (1-7) modulation
method, and the length of a 2T pit, which is the shortest, was 0.2
.mu.m, while in the radial direction, the interval for the pits was
0.4 .mu.m. An optical head having a wavelength of 405 nm and a
numerical aperture (NA) of 0.65 for an object lens was employed to
reproduce data recorded on the ROM disk at a linear speed of 6.6
m/s and at a reproduction power of 0.5 mW (in this case, a clock
cycle T is 15.15 ns). Then, the reproduced waveform was equalized
to PR(12221) to measure the bit error rate bER and the signal
quality evaluation value S. To measure the bit error rate bER, the
original data recorded on the optical disk were compared with the
binary data obtained through Viterbi decoding; and to measure the
signal equality evaluation value S, 10.sup.5 values of reproduced
waveform obtained for each T after PR equalization and the binary
data obtained by Viterbi decoding were employed.
The (1-7) modulated codes are codes having a limitation d>1, and
relative to PR(12221), the Euclid distance is reduced in accordance
with the tap coefficients .epsilon..sub.j shown in Table 1. A set
of tap coefficients .epsilon..sub.j discriminated by pattern 1 in
Table 1, i.e., (1 2 2 2 1), is an error vector that is determined
by two data strings that tend to be erroneously discriminated in
Viterbi detection, e.g., b.sub.k: (1 1 1 1 -1 -1 . . . ) and
c.sub.k: (-1 1 1 1 -1-1 . . . ), and partial response
characteristic (1 2 2 2 1). As for the sets of tap coefficients
.epsilon..sub.i for pattern 2 and the following patterns in Table
1, the number of 0s inserted between (1, 2, 1) and (-1, -2, -1) and
the upper limit number of 0s inserted between (1, 2, 1) and (1, 2,
1) are determined by the upper limit repetitions of the mark/space
for 2T (the upper limit repetitions of a data string of +1 +1 -1
-1, e.g., -1 -1 -1 +1 +1 -1 -1 -1 is counted as one repetition and
-1 -1 -1 +1 +1 -1 -1+1+1+1 is counted as two repetitions). That is,
when the upper limit repetitions of the mark/space for 2T is
(2n+1), the maximum number (4n+1) of 0s are inserted between (1, 2,
1) and (-1, -2, -1). When the upper limit repetitions for the
mark/space for 2T is (2n+2), the maximum number (4n+3) of 0s are
inserted between (1, 2, 1) and (1, 2, 1). Therefore, the signal
quality evaluation value S need only be calculated while taking
into account the pattern up to the upper limit. In Table 1, the tap
coefficients .epsilon..sub.i, up to a maximum of five repetitions
for 2T, are shown, and in this embodiment, 2T was also repeated
five times at the maximum for the pattern that was recorded, using
embossing, on the optical data recording medium.
A pattern having a polarity opposite to that of the pattern shown
in Table 1 (for example, (-1 -2 -2 -2 -1) relative to pattern 1)
may be employed; however, since the obtained signal quality
evaluation value $ is the same, only a pattern having a
predetermined polarity must be taken into account (either polarity
can be determined, but there is no need to evaluate both). In this
embodiment, fourteen taps were fixed, which corresponds to the
number of taps required for the calculation of a sixth signal
quality evaluation value S.sub.6; and to calculate, for example,
the first signal quality evaluation value S.sub.1 for pattern 1,
the fifth and following tap coefficients were regarded as being
0.
TABLE-US-00001 TABLE 1 Example sets of .epsilon..sub.i PATTERNj
.epsilon..sup.i .SIGMA.e.sub.i.sup.2 1 12221 14 2 1210-1-2-1 12 3
121000121 12 4 12100000-1-2-1 12 5 1210000000121 12 6
121000000000-1-2-1 12
Table 2 shows the signal equality evaluation values S measured
while changing a tilt. The j-th signal quality evaluation value
S.sub.1 represents the one obtained for pattern j in Table 1.
TABLE-US-00002 TABLE 2 Relationship between a tilt and S TILT
(DEGREES) S1 S2 S3 S4 S5 S6 -0.3 14 14.5 14.4 14.7 14.7 15 -0.2 15
15.4 15.4 15.9 15.9 16.2 -0.1 16.5 17.5 17.6 18 17.9 18.3 0 17 18
18.1 18.5 18.5 18.8 0.1 16.7 17.1 17.1 17.5 17.6 17.8 0.2 14.8 15.3
15.3 15.8 15.7 16 0.3 13.8 14.2 14.1 14.6 14.6 14.9
As is shown in Table 2, the signal quality evaluation value (first
signal quality evaluation value) S.sub.1 for pattern 1 was the
smallest under all the tilt conditions in the embodiment. FIG. 4 is
a graph showing the relationship between the smallest signal
quality evaluation value S under each tilt condition and the bit
error rate bER. As is apparent from FIG. 4, an obvious correlation
is established between the signal quality evaluation value S and
the bit error rate bER.
In this embodiment, the original data recorded in advance on the
optical disk have been employed to measure the bit rate error bER.
Since it is usually difficult for the original data recorded on a
read-only disk to be accurately known, the measurement of the bit
error rate bER is almost disabled, and it is, therefore, very
difficult for the conduct of the reproduction to be adjusted by
using the bit error rate bER as an index. However, even in such a
case, as is explained in this embodiment, when the reproduction
condition is adjusted by using the signal quality evaluation value
S as an index, data recorded on the optical disk can be stably
reproduced.
As is apparent from Table 2, under each tilt condition the fourth
signal quality evaluation value S.sub.4 or the sixth signal quality
evaluation value S.sub.6 is greater than the second signal quality
evaluation value S.sub.2, and the fifth signal quality evaluation
value S.sub.5 is greater than the third signal quality evaluation
value S.sub.3. Thus, only patterns 2 and 3 in Table 1 may be
employed, as patterns that provide .SIGMA..epsilon..sub.i.sup.2=12,
to simplify the evaluation of the signal quality evaluation value
S.
It should be noted that there are patterns (1 2 1 0 -1 -1 1 1 0 -1
-2 -1) and (1 2 1 0 0 0 1 1 -1 -1 0 1 2 1) as example patterns for
tap coefficients .epsilon..sub.i that provide the third smallest
.SIGMA..epsilon..sub.i.sup.2, i.e., .SIGMA..epsilon..sub.i2=16.
Although not shown in Table 2, the signal quality evaluation values
S obtained for these patterns are not smaller than the signal
evaluation value S obtained for a pattern that provides
.SIGMA..epsilon..sub.i.sup.2=12 or 14. Therefore, the evaluation of
the signal quality evaluation value S is not always necessary for
the pattern that provides .SIGMA..epsilon..sub.i.sup.2=16.
[Second Embodiment]
The relationship between the bit error rate bER and the signal
quality evaluation value S was examined by using a phase change
optical disk formed on a 0.6 mm thick polycarbonate substrate,
wherein the pitches of guide grooves were 0.68 .mu.m. For the
evaluation, while the phase change optical disk was rotated at a
linear speed of 5.6 m/s, an optical head having a wavelength of 405
nm and NA=0.65 was employed to record and reproduce random data
that were obtained through (1-7) modulation at a clock frequency of
64.6 MHz (clock cycle of 15.48 ns). As well as in the first
embodiment, the PR equalization was PR(12221). The random data were
overwritten twenty times, using a recording power of 6 mW and an
erasing power of 2.5 mW, and then, the signal quality evaluation
value S and the bit error rate bER were measured while the focus
offset value was changed. As well as in the first embodiment, the
recording was performed for up to a maximum five repetitions of 2T,
and the first to sixth signal quality evaluation values S.sub.1 to
S.sub.6 were calculated by employing 10.sup.5 values of reproduced
waveforms obtained for each clock after the PR equalization and the
binary data obtained by Viterbi decoding.
TABLE-US-00003 TABLE 3 Relationship between defocusing and S
DEFOCUSING (.mu.m) S1 S2 S3 S4 S5 S6 -0.3 12 11.5 11.8 11.8 12 12
-0.2 15 14.4 14 14.7 14.3 15.1 -0.1 17 16.4 16.1 16.8 16.5 17 0 18
17.5 17 18 17.4 18.3 0.1 16.9 16.3 16 16.9 16.3 17.2 0.2 14.8 14.5
13.8 15 15 15.4 0.3 11.8 11.5 12 11.9 11.9 12.2
As is shown in FIG. 3, under each condition in this embodiment the
signal quality evaluation value S for pattern 2 or 3 was the
smallest. FIG. 5 is a graph showing the relationship between the
smallest signal quality evaluation value S and the bit error rate
bER under each defocusing condition. As is apparent from FIG. 5 as
well as from FIG. 4, there is an obvious correlation between the
signal quality evaluation value S and the bit error rate bER, and
this correlation is the same as that in FIG. 1. It was also
confirmed that, when the smallest signal quality evaluation value S
was equal to or greater than 12, the bit error rate bER was equal
to or greater than 3.times.10.sup.-4. The bit error rate bER
3.times.10.sup.-4 is an index value with which error correction is
enabled by using an ECC (error correcting code), such as
Reed-Solomon code, to obtain a signal at a level that presents no
problem when put to practical use. Therefore, when the reproduction
condition (a tilt or defocusing) is adjusted to obtain the smallest
signal quality evaluation value of 12 or greater, the optical disk
reproduction apparatus can be stably operated.
Since the quality of the reproduced signal may be deteriorated due
to a disturbance, it is preferable that, when the reproduction
condition is adjusted, the bit error rate bER be set lower, by
about one digit, than the threshold value for the stable operation
of the apparatus. To obtain this bit error rate bER, as is apparent
from FIG. 5, the signal quality evaluation value S need only be
equal to or greater than 15, Therefore, it is preferable that the
reproduction condition be adjusted to satisfy a condition wherein
the smallest signal quality evaluation value S is equal to or
greater than 12, and more preferably, is equal to or greater than
15.
As is apparent from Table 3, under each defocusing condition the
fourth signal quality evaluation value S.sub.4 or the sixth signal
quality evaluation value S.sub.6 is greater than the second signal
quality evaluation value S.sub.2, and the fifth signal quality
evaluation value S.sub.5 is greater than the third signal quality
evaluation value S.sub.3. Therefore, only patterns 2 and 3 in Table
1 may be employed, as patterns that provide
.SIGMA..epsilon..sub.i.sup.2=12, to simplify the evaluation of the
signal quality evaluation value S.
As is described above, the recording/reproduction condition can be
adjusted based on the signal quality evaluation value S. However,
as is shown in the first and the second embodiments, since the
smallest signal quality evaluation value S is either the first
signal quality evaluation value S.sub.1, the second signal quality
evaluation value S.sub.2 or the third signal quality evaluation
value S.sub.3, the adjustment of the recording/reproduction
condition can be simplified by examining these three signal quality
evaluation values S.
FIG. 9 is a diagram showing an example of an optical data
recording/reproduction apparatus that includes a function for
adjusting a recording condition or a reproduction condition based
on the signal quality evaluation value S. The quality of a
reproduced signal that is read by an optical head 13 is determined
based on the signal quality evaluation value S obtained by a signal
quality evaluation unit 12. While the recording/reproduction
condition, e.g., a tilt or defocusing or recording power, is
changed by a recording/reproduction condition adjustment unit 14,
the signal quality evaluation value S is calculated for each
condition, and the condition whereunder the signal quality
evaluation value S reaches the maximum, or the
recording/reproduction condition whereunder the signal quality
evaluation value S reaches a constant value (e.g., is equal to or
greater than the 12 described above), is found. As a result, the
data recording/reproduction operation can be performed under a
stabilized condition.
[Third Embodiment]
A third embodiment for calculating the signal quality evaluation
value S will now be described.
By employing
.SIGMA.(v.sub.k+2v.sub.k+1+2v.sub.k+2+2v.sub.k+3+v.sub.k+4).sup.2=N.times-
.(14R.sub.0+24R.sub.1+16R.sub.2+8R.sub.3+2R.sub.4), wherein
R.sub.i=.SIGMA.v.sub.kv.sub.k+j/N is defined, the first signal
quality evaluation value S.sub.1 for the first or the second
embodiment can also be represented by the following equation
(12):
.times..times..times..times..times..times..times..times..times..times..ti-
mes. ##EQU00009## Similarly, the second signal quality evaluation
value S.sub.2 and the third signal quality evaluation value S.sub.3
can also be represented by the following equations (13) and
(14):
.times..times..times..times..times..times..times..times..times..times..ti-
mes..times..times..times..times..times..times..times..times..times..times.-
.times..times..times..times..times..times..times..times..times.
##EQU00010##
R.sub.i corresponds to the autocorrelation of an qualization error,
and it is understood that when the values other than R.sub.0 are 0,
the characteristic of the equalization error is white, The same
calculation can be performed for the signal quality evaluation
values S other than the first, the second and the third signal
quality evaluation values S.sub.1, S.sub.2 and S.sub.3 in Table 1,
or for a set of arbitrary tap coefficients .epsilon..sub.i, so that
the signal quality evaluation value S can be represented as the
auto-correlation function.
FIG. 6 is a diagram showing a configuration example for the
functional block of a signal quality evaluation unit 12A for
calculating the signal quality evaluation value S expressed as the
auto-correlation function. Unlike in FIG. 3, the auto-correlations
for equalization errors are calculated and predetermined weighting
(multiplication of coefficient .beta..sub.i by multipliers 61) is
performed for these results, and the resultant values are added by
an adder 25. To calculate the first signal quality evaluation value
S.sub.1, for example, the coefficients .beta..sub.1 used by the
multiplier 61 need only be set as .beta..sub.0=14, .beta..sub.1=24,
.beta..sub.2=16, .beta..sub.3=8 and .beta..sub.4=2 when the
numerator for the signal quality evaluation value S.sub.1 is
14.times.14. When the numerator for the signal quality evaluation
value S.sub.1 is 14, the coefficients .beta..sub.i need only be set
as .beta..sub.0=14/14, .beta..sub.1=24/14, .beta..sub.2=16/14,
.beta..sub.3=8/14 and .beta..sub.4=2/14. The number of the
multipliers 61 in FIG. 6 is varied, depending on the j-th signal
quality evaluation value S.sub.j to be calculated, and nine
multipliers are required to obtain the third signal quality
evaluation value S.sub.3. In this case also, when depending on the
pattern changing the number of multipliers is complex, only a
satisfactorily large number (15 to 20) of multipliers need be
prepared, and only the coefficients of the multipliers that are not
used for calculation need be set to 0. In FIG. 6, 23-4 is a
multiplier and 23-5 is a mean average calculation unit. The mean
average calculation units 23-5 may be implemented by, for example,
low-pass filters or digital operation units.
Further, in FIG. 6, the acquisition of the mean average may not be
performed (no division is performed using the total number N of
samples). In this case, simply, only the sum described above must
be obtained, and only the numerator of the signal quality
evaluation value S (12 or 14 in this embodiment) must be multiplied
by N.
The measurement process as performed for the first and the second
embodiment was conducted with the configuration in FIG. 6, and it
was confirmed that the same signal quality evaluation value S was
acquired as was obtained in the first and the second
embodiments.
When it is known in advance that the characteristic of the
equalization error is absolutely white, or nearly white, only
R.sub.0 may be calculated.
As is described above, according to the reproduction method using
PR(12221) equalization, predetermined addition and multiplication
are preformed for the auto-correlation of an equalization error
v.sub.m for each clock time, and the first, second and third signal
quality evaluation values S.sub.1, S.sub.2 and S.sub.3 are obtained
using the following equations (15) to (17). Then, the smallest
value is examined to evaluate the quality of the reproduced,
equalized signal.
.sigma..times..sigma..times..times..times..times..times..times..sigma..ti-
mes..sigma..times..times..times..times..times..times..times..sigma..times.-
.sigma..times..times..times..times..times..times..times..times.
##EQU00011## [Fourth Embodiment]
The relationship between the total number of sampling values for
reproduced waveforms and the signal quality evaluation value S was
examined by using the read-only optical disk described in the first
embodiment. The optical head and the measurement conditions, such
as the linear speed and the PR equalization, were the same as those
for the first embodiment, except for the total number of sampling
values. It should be noted that in this embodiment the measurement
was conducted under a condition for which the tilt degree was 0.
FIG. 7 is a graph showing the smallest signal evaluation value S
(the first signal quality evaluation value S.sub.1 in this
embodiment) that was obtained by several measurement operations
performed under each condition for which the total number of
samples was changed. As is shown in FIG. 7, when the number of
samples is smaller than 10000, the signal quality evaluation values
S are distributed across a wide range, which means that adjusting
the reproduction condition is difficult. In order to adjust the
reproduction condition by using the signal quality evaluation value
S as an index, at least 10000 samples are required.
Further, to measure the signal quality evaluation value S in
consonance with the actual disk format, it is convenient for the
measurement to be conducted for each ECC block unit, at the least.
For example, for 64 Kbit ECC blocks, samples for 786432 bits are
employed, and this is satisfactory for the accurate measurement of
the signal quality evaluation value S.
[Fifth Embodiment]
The relationship between the number of sample reproduced waveforms
and the signal quality evaluation value S was examined by using the
phase change optical disk described in the second embodiment. The
optical head and the measurement conditions, such as the linear
speed and the PR equalization, were the same as those for the
second embodiment, except for the total number of samples. It
should be noted that in this embodiment the measurement was
performed with a focus offset of 0. While the total number of
samples was changed, the signal quality evaluation value S (in this
case, the third signal quality evaluation value S.sub.3, which is
the smallest of all) was measured by performing several
measurements under each condition. As a result, as well as in the
fifth embodiment, it was confirmed that the signal quality
evaluation values S were distributed across a wide range when the
number of samples was smaller than 10000. Therefore, to adjust the
reproduction condition using the signal quality evaluation value S
as an index, at least 10000 samples are required.
[Sixth Embodiment]
FIG. 8 is a diagram showing an example functional block for
evaluating the quality of a signal when recorded data are known.
While to calculate an equalization error only the functional block
is shown in FIG. 8, the configuration in FIG. 3 or 6 can be
employed for the signal quality evaluation operation after the
equalization error has been obtained. When the signal quality
evaluation value S is to be measured for data that are recorded by
a recording/reproduction apparatus, instead of employing data
output by a Viterbi decoder, the recorded data that are already
known can be used as reference data. Further, for a reproduction
apparatus, since the same data are always recorded at a
predetermined location on an optical disk, when data are already
known, the data are recorded in the internal memory of the
reproduction apparatus so that they can be employed as reference
data for evaluating the signal quality evaluation value S.
[Seventh Embodiment]
When PR(12221) and record symbol of d>1 are employed, the signal
quality evaluation value S must only be measured for patterns shown
in Table 4. Further, to simplify the process, the signal quality
evaluation values S may be measured only for patterns 1 and 2.
TABLE-US-00004 TABLE 4 Example sets of .epsilon..sub.i PATTERN
.epsilon..sub.i .SIGMA.e.sub.i.sup.2 1 1221 10 2 121-1-2-1 12 3
121-1-1121 14
The combination of (1-7) modulation and PR(12221) have been mainly
employed for the explanation of the embodiments. However, a
combination of another PR method and another modulation symbol may
be employed, and the value S provided, according to the present
invention, as the established form can be calculated, so that the
evaluation of the signal quality and the adjustment of the
recording/reproduction condition are enabled even when the
recording density is too high to measure jitter.
[First Modification]
An explanation has been given for the embodiments wherein the
signal quality is evaluated based on an equalization error for each
clock time, without requiring that an erroneously determined data
string be determined.
However, it is also possible for a predetermined data string that
has been erroneously determined to be determined, and for an
equalization error for this data string to be employed for
evaluating the signal quality provided by an optical disk. In this
case, only a determination unit 71 in FIG. 10 need be provided for
the signal quality evaluation unit in FIG. 3 or 6, and only an
equalization error need be obtained for a data pattern that has
been erroneously determined. The determination unit 71 in FIG. 10
has a function for outputting an equalization error to the
succeeding process block (namely, mean average calculation units
23-6) only for a pattern that has been predetermined.
As is described in the first embodiment, according to the
reproduction method that uses PR(12221), an erroneous detection
tends to occur for a set of data strings that provides
.SIGMA..epsilon..sub.i.sup.2=12 or 14. For example, the set of data
strings that provides 14 is a set of data strings that differ by
only one bit, such as (x 1 1 1 -1 -1 x) and (x 1 1 -1 -1 -1 x) (x
is either 1 or -1). When equation (19) is employed for the
equalization error for the data string (x 1 1 1 -1 -1x) and the
equalization error for the data string (x 1 1 -1 -1 -1 x), the
first signal quality evaluation value S.sub.1 can be obtained. It
should be noted that the data strings used for calculating the
first signal quality evaluation value S.sub.1 are not limited to
these two strings, and all other data strings that differ by only
on bit can be employed.
Furthermore, a set of data strings wherein two bits located at a
distance represented by the time 2T are different, such as (x x 1 1
1 -1 -1 1 x) and (x x 1 1 -1 -1 1 1 x), is an example set that
provides .SIGMA..epsilon..sub.i.sup.2 and that is used to calculate
the second signal quality evaluation value S.sub.2. As well as for
the first signal quality evaluation value S.sub.1, equation (20)
can be employed for the equalization error for data string (x x 1 1
1 -1 -1 1 x) and the equalization error for data string (x x 1 1 -1
-1 1 1 x), so that the second signal quality evaluation value
S.sub.2 can be calculated. It should be noted that the data strings
used for calculating the second signal quality evaluation value
S.sub.2 are not limited to these two, and all other data strings
wherein two bits located at a distance represented by the time 2T
are different can be employed.
In addition, a set of data strings wherein two bits located at a
distance represented by the time 2T are continuously different,
such as (x 1 1 1 -1 -1 1 1 -1 -1 1 1 x) and (x 1 1 -1 -1 1 1 -1 -1
1 1 1 x), is an example set used to calculate the third signal
quality evaluation value S.sub.3.
In this embodiment, the data strings used to calculate the first,
the second and the third signal quality evaluation values S.sub.1,
S.sub.2 and S.sub.3 are recorded in advance on the determination
unit 71, and only when these data strings have been received by the
determination unit 71 is the calculation performed for the signal
quality evaluation value S.
FIG. 10 is a diagram showing an example wherein a signal is
transmitted by a target signal generator 21 to the determination
unit 71. Instead, however, a signal output by a discriminator (a
Viterbi decoder) 20 may be transmitted to the determination unit
71.
[Second Modification]
FIG. 9 is a diagram showing an example optical data
recording/reproduction apparatus that comprises the
recording/reproduction condition adjustment unit 14, which has a
function for adjusting one or both of a recording condition and a
reproduction condition based on the signal quality evaluation value
S obtained by the signal quality evaluation unit 12.
When the recording/reproduction condition adjustment unit 14
adjusts a recording/reproduction condition, such as a tilt, focus
or a recording power, the optical head 13 reads a signal based on
the recording/reproduction condition adjusted by the
recording/reproduction condition adjustment unit 14. The signal
quality evaluation unit 14 then calculates the signal quality
evaluation value S for the signal read by the optical head 13.
Thereafter, the recording/reproduction condition adjustment unit 14
receives the obtained signal quality evaluation value S.
Next, while changing the recording/reproduction condition, the
recording/reproduction condition adjustment unit 14 searches for
the optimal recording/reproduction condition, or for one that, at
the least, is at a specific satisfactory level, so that the signal
quality evaluation value S obtained is the optimum or equals a
predetermined value (e.g., is equal to or greater than the 12
described above). The optical head 13 performs data
recording/reproduction for the optical disk based on the optimum or
a satisfactory recording/reproduction condition that is determined
by the recording/reproduction condition adjustment unit 14. Through
this processing, since the recording/reproduction condition
adjustment unit 14 can determine the optimum or a preferable
recording condition or reproduction condition based on, as an
index, the signal quality evaluation value S that is obtained by
the signal quality evaluation unit 12, the condition under which
data recording/reproduction is performed by the optical data
recording/reproduction apparatus can be stabilized.
By using the method and apparatus of the present invention, the
quality of a signal recorded on an optical data recording medium
can be evaluated at a high recording density under which the
evaluation of the signal quality using jitter is disabled. Further,
the recording/reproduction condition can be optimized by using, as
an index, the signal quality defined in the invention.
* * * * *